proposal

Do you make any cryptographically secure random number generators? I have a personal interest in Fortuna, but am not willing to buy a product that claims to make cryptographically secure random numbers, and does not.

I'm fairly certain nobody wants ran() in general purpose programming languages to use Fortuna directly, so we'll have to live with bad gene- ral purpose random number generators.

Best wishes, and happy random number generation!

Neither is remotely true. Donald Trump has made it more expensive for me to fill my car's gas tank, but not to an extent that I actually notice.

You post a lot of nonsense on sci.electronics.design (and always have) which gives me stuff to be rude about when nothing more consequential comes up. In a sense it is a public service - I get to exercise my throwing arm at regular intervals.

No. A really good random number generator would start with a physics-based source, probably Johnson noise, and scramble that. We just want to simulate random vibration and jitter.

Some people don't want to go off-chip, so use a bunch of bad ring oscillators.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Fair enough. For cryptographical purposes, people need to be aware of /three letter agencies/ subverting physics based sources. This has been mentioned in books by Schneier, so I presume you know all about it, and I won't have to expand further on the subject.

I believe I have somewhere an external noise generator. Forgive me if I don't recall exactly what, as that project has been dormant for some time. It's a ready made /click/ for my PIC32 development board, if I recall correctly. So please don't consider myself as one of those peo- ple.

Have you ever experimented with audio input as a physics based source? I believe you won't even need a microphone, as you should be able to just plug a wire in the mini-jack audio receptor, and use that like an antenna.

Best wishes, and happy random noise generation!

No, I didn't know. How can one subvert Johnson or zener noise? One could easily have a dozen of them and scramble hard.

That could have bad statistics. A real random signal should have no statistics!

Of course it has statistics. Every number should come up equally often and to prove it you have to count up every time a particular number has appeared.

Of course the expected deviation from the predictable total is the square root of the number of observations counted of that particular number so you have to count up a lot of examples of a lot of different numbers to have much confidence that the source is random.

Pseudo=random binary generators are a lot less trouble, if totally predictable.

I don't think you can get a /real random signal/ with a physical input of some kind. And to adjust for the bad statistics -- for now, let's just assume so -- you can fix that on the software end.

Read every /word/ from the ADC [1], and keep only the least significant bit, discard the rest. After thirty two words from the ADC you have a single integer for the digital random number generator.

Of course, that might not be good enough, nor fast enough, for the app- lication, and you can mix that into a more complex R.N.G. with a /linear feedback shift register/ -- if I recall my nomenclature correctly -- or at the deep end of the pool, Fortuna.

So, you think this avenue of noise generation is worth exploring?

I'd have a number of channels. Each a zener diode, amplifier, highpass filter, and a comparator. Feed the comparators into an FPGA and scramble them digitally.

A zener will make ballpark 300 nV/rthz of pretty white noise. That's easier to process than resistor Johnson noise.

I suppose that there must be devices, PCIe cards or USB gadgets, that generate random numbers.

Maybe one could sell memory sticks full of guaranteed random numbers with the understanding that users would use the list sequentially, only once, and buy another when it's all used up. A one-time pad.

YOU HAVE NO CLUE You did not even look a the code, if you did you did not understand it, do you even now C? What happens if NeswFleX exits with and error message is that your can restart in verbose mode, IF YOU HAD LOOKED AT THE CODE you would have noticed, (see this example from articles.c):

struct article *install_article_in_position(\ char *name, char *subject, char *references, int *follow) { int i; int no_match; struct article *pa, *pprev, *plast, *plastmatch, *pnew, *pnext; struct article *lookup_article(); char *test_string; char *ptr1, *ptr2;

if(debug_flag) { fprintf(stdout,\ "install_article_in_position():\nname=%s\nsubject=%s\nreferences=%s\n",\ name, subject, references); } /* argument check */ if(! name) return 0; if(! subject) return 0; if(! references) return 0;

In verbose mode (debug flag set on command line) and you start from a terminal, then you see the arguments to every function displayed on your screen. If it exited on some error found, then the last line you see is the argument of the function that caused the problem. Fault finding in 2 seconds. Noting to do with reboots IDIOT!!!!

The way I see you and your endless sulles babble is a man? in a wheelchair powered by batteries with an old MS windos PC babbling about everything he has zero experience with, waisting people's time because he cannot - or does not - want to learn even the C programming langue or Unix or basic electronics. And then when cornered you refer to insults.

WARNING TO ALL USERS,

DO NOT BOTHER WITH DON.Y IT IS A WASTE OF TIME.

We did one of those for a company called Qrypt a few years back. It was based on shot noise, and worked great, but I don’t think they wound up using it for anything.

Cheers

Phil Hobbs

Shot would be good too. Johnson noise would have a lot of 1/f after amplifying, but that could be filtered.

I remember a tube-based super-regenerative receiver circuit generating many volts of random noise. The physics is interesting, really chaotic.

I did a junior high science project on that and won a medal at the state science fair.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Shoot the messenger, yes? Really? I have a habit of mouthing off and getting things wrong? Let's read the f****ng Wikipedia for a change, you f****ng curmudgeon.

And you can read more about it here.

I was only mildly wrong. It /is/ license free to own the gun. However, to buy the powder and operate it, it requires a license.

I have chosen to completely ignore all your other points, because you seem to be completely free to misrepresent /facts/ in your arguments, while you expect me to be /correct all the time/. So f*ck off!

Thank you. A zener diode is a common suggestion for /real random data/, and is a good choice. I just want to let other members of sci.electron- ics.design know that for cryptographical purposes, it isn't necessarily the best choice. I do not have a ready made reference to a lab experi- ment where a zener diode has been subverted by an adversary, but perhaps you know enough about electronics, to explain how that could be done?

I'll quote a paragraph from /Cryptography Engineering/, by Niels Fergu- son, Bruce Schneier, and Tadayoshi Kohno.

Now the question becomes, how does one check if a zener diode is broken? Is that something that can be added to the circuitry? I'll presume that when such a breakage is detected, the circuitry can trivially notify the host C.P.U., perhaps with a non-maskable interrupt.

I'll add a cautionary tale for reference, since /some people/ think I'm hallucinating all this stuff. This as /Advanced Magickal Dragons/, Zen

For non-cryptographic usage, you can just download up to 16kB from ran- dom.org, so if you mean a physical R.N.G., then I suspect the market is very slim. For cryptographical purposes, an external Fortuna might be worth exploring, especially if your market includes something regulated like finance or military.

I have chosen not to see what's available for sale online, before sub- mitting this followup.

You'll be competing with free downloads from random.org, so you must have something more than that, I'm afraid.

Best wishes, and happy zener diodes!

Seems perfectly in keeping with the line of this thread, which is about nut-cases who like to own guns.

I used to be able to read German. Learning to speak Dutch wrecked my capacity to say anything in German - it always came out as Dutch - but helped my reading comprehension a bit

My guess is that this has to do with local militias, which have persisted long after they were all that useful. Some still rely on cross-bows, but muzzle-loaders must present the same kind of appeal to the traditionally minded.

The American second amendment ""A well regulated Militia, being necessary to the security of a free State, the right of the people to keep and bear Arms, shall not be infringed" is based on exactly the same kind of nostalgia.

Local militias used to be socially important - if you were rich enough to own gun you got to go out and fraternise with other rich people.

The famous Dutch "Nightwatch" painting is a group portrait of just such a group.

It hasn't got much to do with keeping guns out of the hands of people who are silly enough to misuse them - or from the public health point of view, out of the hands of people who might go nuts and misuse them.

At least if the guns owners form some kind of nostalgic militia, they will get together often enough that the body of the group can be expected to notice when one of their number starts to go off the rails,as you seem to be doing.

Plus, what's good is that mentally ill never have any need to defend themselves, because they project a positive aura to a distance of 10'.

We had a thread here, many years on getting a zener diode to oscillate (actually an avalanche diode - any device sold as a zener diode that breaks down above about 6v is actually an avalanche diode). One of my work colleagues had claimed that he had managed to do it many years ago (when he was an apprentice at the UK Royal Radar Establishment at Malvern).

The problem with avalanche diodes is that the avalanche region inside the diode is really small, and if the avalanche current is low enough a single charge carrier can drift through without starting an avalanche, and you have to wait until random fluctuations in the substrate thrown up another. The random current becomes a series of bursts of noise, and if the surrounding wiring is resonant at a convenient frequency you can get more of that frequency than you'd get from a truly random source.

You have to be pretty unlucky to get that by accident.

On 9/25/2026 12:05 AM, Bill Sloman wrote: > On 24/09/2026 7:39 pm, Johann 'Myrkraverk' Oskarsson wrote: >> I'll quote a paragraph from /Cryptography Engineering/, by Niels Fergu- >> son, Bruce Schneier, and Tadayoshi Kohno. >>

Thank you. I found the explanation on Wikipedia confusing. Do you know if the book referenced,

Of course, it's always best to let gun ownership to criminals. They never do anything wrong.

Iceland used to be a gun-free country -- in the sense gun ownership was limited to farmers and hunters -- but now teenagers getting into fights are frequently pulling up guns, at least according to news coverage last time I was there.

In any case, your argument against gun ownership kind of falls down when you start to think about the criminal side of things. Full disclosure, I do not own a gun, nor do I feel the need to.

So, why do you feel the need to rant against gun ownership? Are you a part of a gang who wants to keep the citizens defenseless?

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics